How-to MySQL Performance Diagnostic Using Percona PMM
summary
Live use of Percona PMM plus the U.S.E. method shows a concrete workflow: dashboards point to CPU saturated by table scans, Query Analytics finds one query that is ~1% of queries but causes ~70% of CPU, row scans hitting ~12 million then falling to ~4 million rows/sec after adding an index, and queries per second jumping from ~148 to ~5,000 once the index finishes. The demo builds indexes, measures IO/CPU/memory, and the fix noticeably works.
AI-written summary · about
chapters (16)
- 0:00Start
- 2:08Diagnose Performance
- 4:38Node Summary Node Overview
- 13:21Custom Data Sources with Pmm
- 15:27Network Saturation
- 18:28Thread Cache
- 23:01Check the Resolution
- 25:53My Sequel Summary
- 33:25Access Patterns
- 37:10Query Analytics
- 43:28Mysql Summary
- 45:53Read Only Transaction Commits
- 48:55Storage Summary
- 49:21Index Size per Row
- 53:40Pending Io
- 1:00:15Note Summary
related talks
29mAdvanced Optimization and Troubleshooting MySQL Using PMM2
27mUsing PMM to identify and troubleshoot problematic MySQL queries
30mWhy we recommend PMM to our clients
41mMySQL Query Optimization Best Practices and Indexing
51mHow to improve PostgreSQL Monitoring and ...
51mResolving MySQL Problems Quickly
21mEasy way of application monitoring
28mThe "411" on PMM